COLLOQUIUM Singular Optimal Strategies for a Stochastic Cost Minimization Problem

advertisement
COLLOQUIUM
Singular Optimal Strategies
for a Stochastic Cost
Minimization Problem
Professor Ananda Weerasinghe
Department of Mathematics
Iowa State University
Abstract
In this talk, we describe a stochastic control
problem arising from a controlled queuing network
in heavy traffic. We use Stochastic differential
equations for our model. To solve this control
problem, first we obtain the Hamilton-JacobiBellman (HJB) equation which governs the
associated value function. Then we analyze the
HJB equation to obtain a smooth solution for it.
This solution leads to a singular optimal strategy
for the cost minimization problem. If the time
permits, we will also describe how this optimal
strategy can be utilized to construct nearly optimal
strategies for a cost minimization problem
associated with a controlled queuing network.
Department of
Mathematics
Friday,
April 22, 2011
4:00 p.m.
204 Morgan Hall
Refreshments will be
served at 3:45 p.m.
Download